Fix logic in writeProperty()

[Jojo-Schmitz]

as isStyled() has been asked for already and lead to leaving this method, asking it again doesn't make any sense. Reinstating a comment from 3.x reveals that this method is supposed to be working differently, so now it does.

This would render the force parameter unneeded, if instead setPropertyFlags(<pid>, PropertyFlags::UNSTYLED) would be used.
I guess the force parameter got introduced as a workaround for that bug?

[cbjeukendrup]

This does not match "properties without style are written if different from default value".
Instead, properties without style are now written always, because they are compared to an invalid PropertyValue, rather than the default value.

Since isStyled is indeed checked at the top of the method, isn't the solution (as far as there is anything to be solved at all) to remove this check completely?

Suggested change

	PropertyValue d = !force && (f != PropertyFlags::NOSTYLE) ? item->propertyDefault(pid) : PropertyValue();
	PropertyValue d = force ? PropertyValue() : item->propertyDefault(pid);

[Jojo-Schmitz]

May well be. But why has there been that check in the first place?

Well, as PropertyFlags::STYLED can't happen, != PropertyFlags::NOSTYLE means `== PropertyFlags::UNSTYLED``, and that is when they do get writte (now)

[cbjeukendrup]

But isn't it the wrong way around now? With your change, for NOSTYLE we don't check the default, and for UNSTYLED we do check the default.

So, shouldn't it be like this?

Suggested change

	PropertyValue d = !force && (f != PropertyFlags::NOSTYLE) ? item->propertyDefault(pid) : PropertyValue();
	PropertyValue d = (force || f == PropertyFlags::UNSTYLED) ? PropertyValue() : item->propertyDefault(pid);

[Jojo-Schmitz]

Yeah, most probably... this inverted logic makes me go dizzy...

[cbjeukendrup]

This commit from the history seems relevant: 7523286

[Jojo-Schmitz]

Certainly. Yet the logic is obviously wrong, that case just can't happen

Indeed 3.x has the same issue, that's how I found out...

[Jojo-Schmitz]

Maybe @jthistle can enlighten us here about the intention of his change?

[jthistle]

I wish I could help, but this was almost 7 years ago. I don't even remember writing this commit. How time flies....

Good luck!!!!

[Jojo-Schmitz]

So back to @cbjeukendrup's simplification, that at least keeps the current behavior